Maryam Sadi

During the extraction of oil and natural gas, a huge amount of the brine wastewater is produced. This wastewater, which termed as produced water, contains different impurities, such as salts, heavy metals, inorganic ions and hydrocarbons. By effective management of produced water, its environmental impacts can be minimized. The gas hydrate-based treatment technology has been introduced as a novel technique to treat produced water and eliminate its impurities for industrial applications. Due to the complexity of wastewater treatment processes and the limitation of traditional models, artificial intelligence-based techniques have been proposed in the recent years to simulate the treatment process of wastewaters. In the present study, experimental measurements and modeling technique were performed to study the hydrated-based desalination process of produced water. For this purpose, hydrate formation experiments of carbon dioxide and compressed natural gas were carried out. The experimental setup used in this research, contains a 300 cm3 reactor that is placed in a cooling medium to control the reactor temperature. A computer system, which equipped with an appropriate data gathering software, was applied to record and collect experimental data during hydrate formation period. After collecting required empirical data, an artificial intelligence-based technique was utilized to simulate hydrate-based desalination process. For this purpose, a boosting tree ensemble model was successfully developed to predict the desalination efficiency of produced water as target value by considering initial salinity and gas equilibrium pressure as input parameters. After model development, wide varieties of graphical and statistical error analysis techniques were applied to assess the smart model performance. The coefficient of determination (R2) and mean absolute percentage error (MAPE) of the proposed model for all experimental data were 0.9938 and 0.71%, respectively, which indicate the excellent agreement between model predictions and experimental data. Based on the obtained results, it can be concluded that the developed tree-based model can be applied by high accuracy in predicting desalination efficiency of produced water through the hydrate-based desalination process. Furthermore, the measured data reveal that the CO2 gas hydrate has a higher average desalination efficiency compared to compressed natural gas. This phenomenon can be attributed to the well-packed structure of CO2 hydrate in contrast to the spongy form of natural gas hydrate. In fact, in a well-packed hydrate structure, less salt was trapped between the hydrate crystals, therefore, the water obtained after hydrate dissociation has a lower salinity.

Co-author/s:

Hamid Ganji, Head of Gas Research Division, Research Institute of Petroleum Industry.

Hajar Fakharian, Researcher, Amir Kabir University of Technology.